Showing papers on "Channel (digital image) published in 1994"

Document image decoding using Markov source models

Abstract: Document image decoding (DID) is a communication theory approach to document image recognition. In DID, a document recognition problem is viewed as consisting of three elements: an image generator, a noisy channel and an image decoder. A document image generator is a Markov source (stochastic finite-state automaton) that combines a message source with an imager. The message source produces a string of symbols, or text, that contains the information to be transmitted. The imager is modeled as a finite-state transducer that converts the 1D message string into an ideal 2D bitmap. The channel transforms the ideal image into a noisy observed image. The decoder estimates the message, given the observed image, by finding the a posteriori most probable path through the combined source and channel models using a Viterbi-like dynamic programming algorithm. The proposed approach is illustrated on the problem of decoding scanned telephone yellow pages to extract names and numbers from the listings. A finite-state model for yellow page columns was constructed and used to decode a database of scanned column images containing about 1100 individual listings. >

Method for rectifying channel errors in a transmitted image signal encoded by classified vector quantization

Abstract: A novel method for concealing or rectifying channel errors present in a decoded image signal which has been transmitted in a compressed form by using a classified vector quantization technique comprises a step for estimating the class for a current block by detecting boundary edges at each of four groups of neighboring pixels; and a step for estimating the representative vector for the current block by calculating a multiple number of side-matching functions and choosing the candidate vector producing the smallest side-matching function as the representative vector for the current block.

Reconstruction of a high-resolution image from multiple-degraded misregistered low-resolution images

Abstract: In applications that demand highly detailed images, it is often not feasible or sometimes possible to acquire images of such high resolution by just using hardware (high precision optics and charge coupled devices). Instead, image processing approaches can be used to construct a high resolution image from multiple, degraded, low resolution images. It is assumed that the low resolution images have been subsampled (thus introducing aliasing) and displaced by sub-pixel shifts with respect to a reference frame. Therefore, the problem can be divided into three sub-problems: registration (estimating the shifts), restoration, and interpolation. None of the methods which appeared in the literature solve the registration and restoration sub-problems simultaneously. This is sub-optimal, since the registration and restoration steps are inter-dependent. Based on previous restoration and identification work using the Expectation-Maximization algorithm, the proposed approach estimates the sub-pixel shifts and conditional mean (restored images) simultaneously. In addition, the registration and restoration sub-problems are cast in a multi-channel framework to take advantage of the cross- channel information. Experimental results show the validity of this method.© (1994) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Multichannel restoration of single channel images using a wavelet-based subband decomposition

Abstract: We present a new matrix vector formulation of a wavelet-based subband decomposition. This formulation allows for the decomposition of both the convolution operator and the signal in the subband domain. With this approach, any single channel linear space-invariant filtering problem can be cast into a multichannel framework. We apply this decomposition to the linear space-invariant image restoration problem and propose a family of multichannel linear minimum mean square error (LMMSE) restoration filters. These filters explicitly incorporate both within and between subband (channel) relations of the decomposed image. Since only within channel stationarity is assumed in the image model, this approach presents a new method for modeling the nonstationarity of images. Experimental results are presented which test the proposed multichannel LMMSE filters. These experiments show that if accurate estimates of the subband statistics are available, the proposed multichannel filters provide major improvements over the traditional single channel filters. >

A numerical approach to the analysis and classification of channel network patterns

Abstract: A large number of samples of visually classified channel networks are taken from some published works. The samples are representative of the following pattern types: dendritic, parallel, rectangular, trellis, and pinnate. By means of computer techniques, essential parameters of elements of these pattern samples are determined. These include lengths, directions, and degrees of curvedness and meandering of channel segments, as well as confluence angles. The parameters are used to determine pattern attributes such as density, texture, parallelism, rectangularity, and bifurcation ratios. Thresholds are generated for the attributes and used in the construction of classification models for the five pattern types studied. The processes and models are coded in a computer program for use in the automated classification of numerically valued channel networks. After classification the image is displayed with each individual network bearing a color which shows the pattern type to which it has been assigned.

Stochastic modeling and estimation of multispectral image data

Abstract: The application of a monochromatic restoration algorithm to each channel within a multispectral image does not result in an estimate which takes into account cross-channel correlation. A non-Gaussian prior model is proposed for multispectral images, using the Gibbs distribution. The density accounts for both spatial (within channel) and spectral (between channel) information. Spatial components use a nonlinear operator to preserve discontinuities within each channel, while spectral components incorporate cross-channel information in the model. The prior density is used in a maximum a posteriori (MAP) estimation algorithm for the restoration of color images. The resulting nonlinear estimates are shown to be quantitatively superior to linear estimates produced by multichannel Wiener and least squares restoration, which implicitly use Gaussian priors. >

Color resolution enhancement by using color camera and methods

Abstract: A color resolution enhancement method and enhanced color cameras (600) with a single CCD detector (606) having attached color filters (604) and using luminance-based gradient interpolation to fill in color channels A color camera, includes an image detector, the detector with sets S 1 , S 2 , S N of pixels for N, an integer greater than 1, where pixels in the set S J detect radiation in a corresponding sampled channel C J for each J in the range 1 to N, a channel reconstructor coupled to an output of the detector The reconstructor includes a luminance generator which combines the sampled channels C 1 , C 2 , C N to a luminance channel, a gradient generator coupled to an output of the luminance generator and which transfers the luminance channel to a gradient channel, and a directional interpolator coupled to an output of the detector and to an output of the gradient generator and which transfers each of the sampled channels C 1 , C 2 , C N to a corresponding reconstructed channel RC 1 , RC 2 , RC N , and a reconstructed channel output

Method for enhancing detail in color signals and circuitry for implementing that method in color video equipment

Abstract: In video signal processing circuitry detail enhancement is done on each of three color channels in response to the detail information, the enhancement of the detail in each color channel being done in reliance on separated high-spatial-frequency information originally contained in that channel, as long as it is sufficiently greater in amplitude than a reference level somewhat above that expected for thermal noise. When the amplitude of the detail information originally contained in a channel is insufficiently above that reference level, the thermal noise is suppressed by subtracting the separated high-spatial-frequency information therefrom. The same filter arrangements are used to separate high-spatial-frequency information for suppressing thermal noise in each color channel as for enhancing the detail in each color channel.

Digital image color compensation with unequal integration periods

Abstract: In the preparation of fluorescence in situ hybridization (FISH) specimens, when multiple DNA probes, each tagged with a different fluorophore, are used in combination, several different cellular structures can be visualized simultaneously. When such specimens are digitized, the different cellular components appear in separate color channels of the multi-channel digital image. Spectral overlap among the emission spectra of the fluorophores, the sensitivity spectra of the image sensor and the passbands of the optical filter set, however, spread the image of each fluorophore across the color channels. The result can be poor isolation of each fluorophore to a separate color channel. Each probe image is thus contaminated by ‘crosstalk’ from the other channels, to a greater or lesser degree. In an earlier paper we presented a processing method that effectively removes the spreading of fluorophore brightness among the color channels, thereby isolating each fluorophore to a single (monochrome) image. That technique assumes the black level is zero, and the integration time is the same for each channel. This paper extends the technique to account for non-zero black level and for unequal integration periods for the different fluorophores.

Document Image Decoding Using

Abstract: Document Image Decoding (DID) is a communica- tion theory approach to document image recognition, patterned after the use of hidden Markov models in speech recognition. In DID, a document recognition problem is viewed as consisting of three elements-an image generator, a noisy channel and an image decoder. A document image generator is a Markov source (stochastic finite-state automaton) that combines a message source with an imager. The message source produces a string of symbols, or text, that contains the information to be transmitted. The imager is modeled as a finite-state transducer that converts the one-dimensional message string into an ideal two-dimensional bitmap. The channel transforms the ideal image into a noisy observed image. The decoder estimates the message, given the observed image, by finding the Q posteriori most probable path through the combined source and channel models using a Viterbi- like dynamic programming algorithm. The proposed approach is illustrated on the problem of decoding scanned telephone yellow pages to extract names and numbers from the listings. A finite-state model for yellow page columns was constructed and used to decode a database of scanned column images containing about 1100 individual listings. Overall, 99.5% of the listings were correctly recognized, with character classification rates of 98% and 99.6%), respectively, for the names and numbers.

Context sensitive color quantization system and method

Abstract: A system and method usable in producing a display image limited to a lesser number of colors from an input image having a greater number of colors. The invention is particularly useful in a computer system utilizing a video monitor capable of simultaneously displaying a selected number of colors stored in a lookup table. The system and method receive data representative of respective colors at a multiplicity of individual pixels in the input image. This input image data is then temporarily stored in an input buffer. An appropriate processor receives the input image data and performs image quantization thereon. Specifically, the processor first determines the selected colors to be included in the display image. Next, the processor determines which of the selected colors respectively correspond to individual pixels of the display image. Unlike the prior art, image quantization algorithms of the present invention selectively account for perceptive blending of colors appearing in the input image in determining the optimum display image. In other words, if individual colors appear in the input image in such a manner that a viewer would perceive a blended color, the invention will tend to display the blended color rather than the individual colors.

Image processor with input buffering to multiple digital signal processors

Abstract: A system architecture is provided that includes an image sensor unit operable in a single channel mode and a dual channel mode. The image sensor unit includes an electronic image sensor comprising a row and column array of pixel elements, wherein the rows of the array having a line length of N pixels. First and second digital signal processing units for processing image data generated by the image sensor unit into color component image data are provided, wherein each of said first and second digital signal processing units has a line length processing capacity less than N pixels. An input buffer, coupled between outputs of the image sensor unit and inputs to the digital signal processing units, simultaneously receives two lines of image data from the image sensor unit in the dual channel mode of operation, and sequentially supplies a first portion of each of the simultaneously received image lines to the first digital signal processing unit and a second portion of each of the simultaneously received image lines to the second digital signal processing unit. An output buffer, coupled to the output of the first and second digital signal processing units, combines the color component image data generated by the first and second digital signal processing units into color component image lines of length N pixels. A control unit controls the operation of the image sensor unit, the input buffer, the first and second digital signal processing units, and the output buffer. A frame store receives and stores the color component image lines generated by the output buffer, and a monitor displays the image lines at a frame rate that is about twice the operating frame rate of the image sensor unit.

A high performance Micro Channel interface for real-time industrial image processing applications

Abstract: Data collection and transfer devices are critical to the performance of any machine vision system. The interface described in this paper collects image data from a color line scan camera and transfers the data obtained into the system memory of a Micro Channel-based host computer. A maximum data transfer rate of 20 Mbytes/sec can be achieved using the DMA capabilities of the camera interface. Programmability of the interface provides flexibility in selection of features such as collected field-of-view, data format, collection method, and camera control. System design is such that the speed of the interface allows for the collection of images intended for real-time processing by any available Micro Channel resident processor(s). >

Multichannel image identification and restoration using the expectation-maximization algorithm

Abstract: Previous work has demonstrated the effectiveness of the expectation-maximization algorithm to restore noisy, degraded single-channel images and simultaneously identify its blur. In addition, a general framework for processing multi-channel images using single-channel techniques has also been developed. This paper combines and extends the two approaches so that simultaneous restoration and blur identification is possible for multi-channel images. However, care must be taken in estimating the blur and the cross-power spectra, which are complex quantities. With this point in mind, explicit equations for simultaneous identification and restoration of noisy, blurred multi-channel images are developed, where the images may have cross-channel degradations. Experimental results are shown which support this multi- channel approach, and are compared with multi-channel Wiener filter results. Independently restoring each channel is also analyzed and compared with multi-channel results.

Digital video signal recording apparatus and method

Abstract: A digital video signal recording apparatus records a video signal expressed by four digital components such that even numbered pixel is expressed by a first type luminance signal component and two color difference signal components, and odd numbered pixel is expressed by a second type luminance signal component. For the 525/60 NTSC system with 3-segment and 4-channel arrangement, one field video signal is stored in three segments. The recording apparatus has a standard channel order generator for generating a standard channel order of a recurring channel pattern. The four digital components are aligned based on the recurring channel pattern so that the first type luminance signal component and two color difference signal components for the even numbered pixel are stored in a same segment by the same channel. As a result, the samples for the same-position pixels are recorded by the same recording head, adjacent samples are recorded by different heads, and error concealment functions effectively.

High-performance infrared filters for the HIRDLS 21-channel focal plane detector array

Abstract: The High Resolution Dynamics Limb Sounder is described, with particular reference to the atmospheric measurements to be made and the rationale behind the measurement strategy. The demands this strategy places on the filters to be used in the instrument and the designs to which this leads to are detailed. A second set of filters at an intermediate image plane to reduce 'Ghost Imaging' is discussed together with their required spectral properties. The spectral characteristics of the primary and secondary filters in each channel are combined together with the spectral response of the detectors and other optical elements to obtain the system response weighted appropriately for the Planck function and atmospheric limb absorption. This method is used to demonstrate whether the out-of-band spectral blocking requirement for a channel is being met, and an example calculation is demonstrated showing how the blocking is built up for a representative channel. Finally, the techniques used to produce filters of the necessary submillimeter sizes together with the testing methods and procedures used to assess the environmental durability and establish space flight quality are discussed.

Scheme for transmitting encoded blocks of video signals at different channel rates

Abstract: Encoded blocks (from 110) of video information are selected (120,130,170) for transmission at a plurality of channel rates (C₁, C₂) based on the corresponding location of the portions of the video image represented by the encoded blocks. Each encoded block is designated an image type which determines its transmitted channel rate. The image types are associated with spatial positions of a video image such that encoded blocks within a first area of the video image are designated a first image type and encoded blocks within a second area of the image are designated a second image type. In general, data blocks representing the central portion of the video image are transmitted at a channel rate having less susceptibility to noise or other degradation, thereby ensuring that such portion can be received in fringe areas.

Multichannel stochastic image models: theory, applications, and implementations

Abstract: The restoration and enhancement of multichannel data, including multispectral images and video sequences, is now feasible, due to recent advances in computing technology. Within each channel, an assumption of stationarity is reasonable. However, spectral and temporal correlations are nonstationary, so that a simple extension of single channel estimation algorithms to multichannel data will not give the best results. Bayesian maximum a posteriori (MAP) estimation requires a model for the data to regularize ill-posed inverse problems such as image restoration and reconstruction. Quite often, an assumption of global smoothness is made for the data, which can be modeled by a Gaussian prior. If noise corrupting the data is Gaussian as well, this results in linear Bayesian estimates containing smooth edges. A non-Gaussian prior is presented in this dissertation which models piece-wise smooth data, and thus discontinuities are more likely to appear in the resulting nonlinear estimate. A number of topics involving Bayesian image restoration are investigated in this dissertation. A window-based Bayesian estimator is developed for use in image noise removal, constraining the amount of data used in estimating each pixel. Restorations using the window-based estimator have the same quantitative and visual quality as Bayesian estimates computed using all observed image data, even for small window sizes. The stochastic image prior is extended to model multispectral data, featuring spectral Markov random field clique functions which incorporate cross-channel correlations between frequency bands. Estimates from the corresponding Bayesian color image restoration algorithm are improved over independent channel estimates of the red, green, and blue color planes. A maximum likelihood technique for automatically estimating non-Gaussian signal and image model parameters directly from noisy and blurred observations is proposed. Numerical simulations confirm that the estimated parameters result in high quality signal reconstructions. Finally, an observation model for low-resolution video sequence data is presented, which accounts for the motion between frames. A Bayesian interpolation algorithm which extracts a high-resolution frame given several frames from a low-resolution video sequence is then described, with significant improvement over single frame interpolation methods.

Verfahren zum Übertragen von Daten

Abstract: An image transmission process allows fixed data rates to be efficiently utilised The image data to be transmitted are sorted into different categories depending on their sensitivity to channel errors and the categories are provided with different error protections and transmitted Information words that contain information on the number of data bits belonging to the category and the type of error protection used precede the individual categories In addition, a synchronisation word is added to the image data of each image At the receiver, the first information words are channel-decoded according to a fixed coding rate and the information stored in the first information words is used for channel-decoding the image data The image data are then image source decoded and outputted

One-dimensional color image sensor

Abstract: PURPOSE: To provide a one-dimensional color image sensor, which can narrow the distance between neighboring sensor elements in the main scanning direction and can perform the high-speed operation. CONSTITUTION: Output lines 24 from the cathodes of three photodiodes 20a-20c for one pixel are commonly connected. The output lines of the corresponding block channel for each block are made common and connected to four output lines 2a-2d, respectively. The light receiving surface of the photodiode can be effectively utilized by forming the output line of a transparent conducting film. A sensor element can perform the high-speed operation especially required when coloring is performed by constituting the sensor element of the diode and TFTs 22a-22c. Another TFT is provided as a switching means on the input side of the TFT of the sensor element. Any of three colors is selected based on a signal applied on the gate electrode and can be outputted. COPYRIGHT: (C)1995,JPO

Video communication system

Abstract: PURPOSE:To attain simultaneous transmission of the entire state and detailed partial image by putting, transmitting and displaying both of an entire image of a comparatively rough scene and a detailed partial image simultaneously. CONSTITUTION:A picture inputted from a wide angle camera 1 is resampled on a logarithmic polar coordinate system at a 1st coordinate conversion section 3 to obtain a wide angle coordinate conversion picture. A visual axis arithmetic operation section 5 calculates a visual axis direction of a standard camera 2 and generates a visual axis signal based on a wide angle coordinate conversion picture. A visual axis signal is inputted to a motor-driven frame 6 and the visual axis of the camera 2 is moved in the direction. An input picture from the camera 2 is similarly re-sampled to obtain a standard coordinate conversion picture. Both the conversion pictures is subject to redundancy suppression processing at a coding section 7 and the result is composed into one channel video signal and it is transmitted. The received video signal is given to a decoding section 8, in which the signal is decomposed into original 2 systems of video signals and visual axis signal at a decoding section 8 and the video signal receives decoding processing. The 2 systems of decoded pictures are converted into orthogonal coordinate system pictures by coordinate conversion sections 9,10 and the wide and standard decoded pictures are obtained.

Focal distance controlling stereoscopic-vision television receiver

Abstract: PURPOSE: To form the stereoscopic-vision television receiver with one channel in accordance with the NTSC standards without use of an eyeglass by using a lens whose focal distance is varied by electrons for all pixels of the television receiver and controlling a virtual vision of the pixcels with respect to all pixels. CONSTITUTION: A stereoscopic vision is obtained basically by controlling a virtual vision appearing in the vertical and horizontal directions by a projection lens or a concave mirror. That is, a focal distance variable lens whose focal distance is varied by electrons is used for all pixels of the television receiver to control a virtual vision of each picture element with respect to all the pixels. For the purpose of it, distance information is to be added to an NTSC signal. The detailed highs and lows of all small pixels are controlled based on the information and one virtual vision is formed through vision forming. Then a side visual sight is obtained through the combinations of large sized concaved and convex lenses. Furthermore, one channel is enough for the purpose, the NTSC stereoscopic-vision television receiver without eyeglass and taking over the existing signal is realized.

Automatic characteristic inspecting and adjusting device for color receiver

Abstract: PURPOSE:To electrically perform the inspection and adjustment of picture quality such as the channel synchronizing frequency, hue and contrast of a color receiver by providing an arithmetic means and a detection means for detecting the synchronizing condition of the color receiver and automatically inspecting the characteristic of the receiver CONSTITUTION:A still test pattern is displayed on the screen of a color receiver 110 to be inspected by a signal generator 109 The image is imaged as two sheets of video after fixed time by an image pickup device 101 and the video signal is stored in memories 104, 105 by an image switching device 103 via an A/D converter 102 A subtraction processing is performed for the image data in a subtracter 106 and the result is stored in a memory 107 A judgment circuit 108 judges that the setting of the channel frequency of the receiver 110 is not normal when the subtraction result is a finite value which is not zero

Immediate interactive reception method for video image

Abstract: PURPOSE:To allow a receiver side to attain still processing of a video image, fast feed, rewinding and slow display or the like during video image distribution by selecting one video image as a reception video image among distributed video images delayed respectively and allowing the receiver side to select the video image. CONSTITUTION:A video image (f) inputted from an input side 11 is copied and distributed at a video image distribution delay discrimination section 12 and in this case, each distributed video image is referred to as f1K (k=0-n). The distributed video image f1K (k=1-n) is delayed at a delay section 13 by a different delay DELTAtK (k=1-n) and each delayed video image f2K (k=1-n) is obtained. Furthermore, the distributed video image f10 is a video image on a video channel as it is. Then each delayed video image f2K is synthesized at a synthesis section 14 as a sub channel of a video channel. That is, since each video image is delayed by various quantities, the selected video image is properly selected by preference of a recipient.

Sparse colour and gre y scale image resto rat ion using a morphological method

Abstract: In an experimental study on colour and grey scale images coveting a range of corruption from 50 to 95%, good restoration was achieved using a new morphological method. The nearest good neighbour (NGN) morphological filter copies grey scale values from the 'good' pixels in a regular manner so that these pixels become seeds for the restoration during successive iterations. A complementary propagation method is also described. In this context, sparse data refers to an image in which a large fraction of the data has been replaced by impulsive noise. The impulse noise may have a high value or range or a zero or n u l l v a l u e . R a n d o m l o s s o f a n i m a g e communication channel will result in a sparse image which may be restored by these methods. Range images or optic flow data may be processed by these methods also. On a face image which had lost 95% of its data, the restored image had a signal to noise ratio of 16dB and all features were clearly discernable. The filter had an 8 pixel neighbourhood and took about 0.5 second per iteration to filter the image on a 386 standard PC.

Correction of influence of alternating voltage coupling in a signal chain of scanning opto-electronic sensors

Abstract: A method for correction of influence of alternating voltage coupling in a signal chain of scanning opto-electronic sensors on radiometric data thereof is carried out by firstly ascertaining, and storing, characteristic times from the time constant of a RC-circuit of the signal chain of the respective spectral channel, a line scanning frequency and a number of image elements per scanline. A preliminary correction in the form of a recursive inverse RC-filtering of the sampling values is then performed for each image element of a channel with its characteristic time, and the preliminary corrected image data are intermediately stored channel-by-channel and line-by-line. In the case of the use of a dark reference source completely independent of external radiation, corrected pixel values are computed immediately and directly on the basis of the preliminarily corrected and intermediately stored image data with the aid of dark reference signals present at the sensor output for each channel.